Typically the prevalence of dental caries is determined by a full mouth examination for each subject. However, limitations of manpower or time may often preclude such a approach. Knutson showed that a screening examination using the presence or absence of caries could be used to predict DMFT scores for children. In this study we investigate how well DMFS scores for subjects under 35 years of age can be estimated by specific models based on age-specific prevalences, using data from the NIDR prevalence surveys (1987 children's survey and the 1985 adult survey). Data for adults older than 35 were not used, since the M component in DMFS is not caries specific for these adults. Knutson in his original model studied the relationship between the age- specific mean DMFT and the age-specific caries prevalence. Here, we considered two models. One is Knuston's model with DMFS replacing DMFT. The second one is the linear regression model. The assumption for the linear regression model is that the prediction of DMFS can be expressed as a linear function of the log of the proportion of caries-free individuals and age. Both Knutson's model and the regression model can be used in both children and young adults separately to describe the relationship between caries severity (DMFS) and caries prevalence. The regression model is easier to interpret and estimate than the nonlinear Knutson model. The regression model also emphasized the need to adjust for age.